Got "Failed to Resolve hostname of sentinel" but no other log to look for reason

[royatanu94]

Hi,
We have valkey deployment where 3 master and 3 sentinel were deployed. Our application was connected to valkey sentinel but after some time communication is broken.
We only see "Failed to Resolve hostname of sentinel x" in sentinel logs, other than that no information is available to debug the issue further.
Please help to debug the issue further.

[royatanu94]

@rueian Any help is really appreciated.

[hieu2102]

Hi, can you provide more details on this?

• what version of valkey are you using
• the series of commands you used to setup the sentinel deployment

From the description, I'm guessing it's because the nodes weren't able to resolve the hostname (translate hostname to IP).

Thanks

[royatanu94]

We have used bitnami deployment yaml and can you share which slack channel i should use to post query?

Can you give any pointers where to look for this issue?

[hieu2102]

Here's the slack for valkey: https://valkey.io/slack

You should post this query in the valkey-users channel there :D, it seems to be one of the more active channels

[hieu2102]

Sorry, but I couldn't replicate your issue.

• I couldn't find the image version 7.2.4 in Bitnami's Docker Hub repo
• when I deploy using image version 7.2.5, I got this error (because at that version, valkey still use master/slave keywords instead of primary/replica

*** FATAL CONFIG FILE ERROR (Version 7.2.5) ***
Reading the configuration file, at line 4
>>> 'primaryauth "XqJuz5IuCK"'

Maybe try upgrading the helm templates and try again, to see if the error still occur?

Personally, I expect the problem is around your k8s cluster DNS server, rather than valkey.

[royatanu94]

Thank You.
Could you please point out what to check in k8s DNS? what can cause it? As initially master and slave were connected but after some day we got that issue

[hieu2102]

Maybe the official k8s documentation could help you: https://kubernetes.io/docs/tasks/administer-cluster/dns-debugging-resolution/

I also see this issue looks similiar to what you described: kubernetes/kubernetes#110550